A Novel Information-Directed Tree-Search Algorithm for RIS Phase Optimization in Massive MIMO

In this paper, we propose a novel Information-Directed Branch-and-Prune (IDBP) algorithm to optimize the discrete phase settings of a passive Reconfigurable Intelligent Surfaces (RIS) to improve the blocked line-of-sight (LOS) link performance in the Massive Multiple-Input Multiple-Output (MaMIMO) wireless networks. The MaMIMO transceivers are assumed to be equipped with a hybrid precoder, combiner, and low-resolution analog-to-digital converters (ADCs). Mean Squared Error (MSE) of the received and combined symbol at the receiver is used as the performance metric. We derive the expression for the Cramer-Rao lower bound (CRLB) of the MSE as a function of the RIS phase shift settings for a given hybrid precoder, combiner, and ADC bits. The proposed IDBP algorithm arrives at the RIS phase settings that minimize the CRLB using an information-theoretic measure that define the pruning rules in the tree traversal. Using the proposed algorithm and with an appropriate design of the hybrid precoder and combiner, it can be shown that the MSE achieves the CRLB. We claim theoretical guarantees for near-optimal MSE performance and the throughput for a blocked LOS link with reduced computational complexity.